24 Cultural Safety

Loubna Belaid and Neil Andersson


A concept originating from nursing sciences and here applied in an innovative way to evaluation, cultural safety refers to an approach aimed at ensuring that the evaluation takes place in a “safe” manner for the stakeholders, and in particular for the minority communities targeted by the intervention under study, i.e. that the evaluation process avoids reproducing mechanisms of domination (aggression, denial of identity, etc.) linked to structural inequalities. To this end, various participatory techniques are used at all stages of the evaluation. Cultural safety is compatible with all types of methods. It contributes to making the evaluation more relevant and useful for stakeholders and will likely increase their self-determination.

Keywords: Mixed methods, participation, indigenous evaluation, culturally responsive evaluation, inequalities, racism, decoloniality, fuzzy cognitive mapping

I. What does this approach consist of?

The cultural safety approach was developed in response to the observation that the evaluation process could reproduce mechanisms of domination linked to structural inequalities, particularly concerning indigenous peoples or in post-colonial contexts. For example, one study assessed the perception of three psychometric scales used to diagnose depression in the Inuit and Mohawk populations of Quebec. The study results showed that the three scales were not culturally safe.  Participants disliked the numerical assessment, the self-report (as opposed to supportive interaction) and the focus on symptoms rather than supportive factors (Gomez Cardona et al. 2021).

Cultural safety aims to produce ‘an environment for people where there is no aggression, challenge or denial of who they are, what they need’ (Williams 1999). A Maori nurse originally developed the concept in response to the racism and discrimination faced by Maoris in healthcare settings (Papps and Ramsden 1996). Culturally safe evaluation aims for stakeholders to feel that their cultures are respected and strengthened by the evaluation.

Cultural safety goes beyond another approach more commonly promoted in evaluation based on cultural sensitivity and competence (culturally responsive evaluation). Indeed, beyond simply paying attention to cultural differences, cultural safety considers the power imbalances, institutional discrimination, racism and colonial relations that can interfere with the design and implementation of services and programmes (Curtis et al. 2019). The concept is thus situated within the spectrum of critical postcolonial theories and aims for social justice. The concept of cultural safety has been extended beyond Maori communities to any group that differs from its care or service providers regarding age, gender, socio-economic status, ethnicity, religion or disability (Smye and Browne 2002).

The concept has attracted the attention of research and evaluation approaches that challenge unidirectional and conventional perspectives centred on the point of view of the person conducting the study, with a minimal contribution or benefit to the participants (Smith 2012; Cram 2016; Katz et al. 2016).  It also calls for a revisiting of concepts, methods, values and evaluation approaches stemming from Western epistemologies and draws attention to the validity of worldviews of indigenous and minority groups (Smith 2012; Belaid et al. 2022)

Cultural safety aims to ensure that evaluation is beneficial and relevant to these communities. It aims to empower them and can contribute to increasing their self-determination (Gollan and Stacey 2021). Cultural safety changes the direction of evaluation by incorporating the participants’ perspective.

Five key principles characterize cultural safety in evaluation (Wilson and Neville 2009; Cameron et al. 2010; Andersson 2018): (i) participation (ii) partnership (iii) ownership (iv) critical reflexivity and (v) protection of identities, beliefs, cultural values and worldview.

  1. Participation: refers to the involvement of stakeholders throughout the evaluation (Cameron et al., 2010). Procedural or symbolic participation should be distinguished from genuine participation. Procedural participation allows for structured stakeholder input at specific stages of the process, for example, structured interviews with key informants. Token participation may involve a paid or unpaid stakeholder ‘representative’ participating in certain evaluation activities. Authentic participation includes co-ownership of the evaluation, active engagement in the analysis of evidence and a role in designing solutions based on the evaluation findings.

  2. Partnership formalizes participation, often evolving into genuine participation as the evaluation unfolds. It is about establishing equitable relationships between the evaluation team and stakeholders, whether they are communities, patients or staff. The evaluation team should clarify the potential scope of these relationships at the outset, strive to maintain them, and allow them to evolve as stakeholders increase their capacity throughout the evaluation (Cameron et al. 2010).

  3. Ownership of the evaluation process, results and governance: cultural safety enables stakeholders to take ownership of the process (Andersson 2018). This can start with circumscribing the purpose of the evaluation within the limits of the funded objectives, or ‘having a voice’ in what is being evaluated and how, and being involved in the evaluation activities, including interpreting the results.

  4. Critical reflexivity: The starting point for culturally safe evaluation is for evaluators to reflect on their values and beliefs, social position, power and privilege (Wilson and Neville 2009; Browne et al. 2016). This involves an awareness of the historical relationship between evaluators and indigenous and minority communities, the history of colonialism, and the systemic racism and discrimination that these communities may still face (Cameron et al. 2010).

  5. Protection strengthens research ethics by protecting indigenous and minority groups from exploitation and reinforcement of negative representations or accounts (Wilson and Neville 2009). This implies that their knowledge, values and epistemologies are also valued alongside Western scientific epistemologies and methods (Cameron et al., 2010). Indigenous communities thus want to see evaluation rooted in their worldview (Belaid et al. 2022).

Several frameworks and guidelines address how to develop a more fair evaluation when involving indigenous and minority groups (Wilson and Neville 2009; Cameron et al. 2010; Gollan and Stacey 2021). Here we present the framework developed by Andersson and colleagues to clarify how cultural safety unfolds at different evaluation stages (Cameron et al. 2010; Andersson 2018).

Formulation of the purpose of the evaluation

Ideally, a culturally safe evaluation emanates from a community request. This increases the relevance of the evaluation, ensuring alignment with community priorities. In practice, many evaluations are commissioned from a problem defined by donors, leaving less room to formulate or rename the purpose of the evaluation.

Culturally safe evaluation chooses to build on the strengths of communities. Dominant social groups often label indigenous and minority groups as ‘at risk’, ‘vulnerable’ or ‘marginalized’. These negative labels do not necessarily reflect how these communities see themselves and significantly reduce the space and conditions for improvement (Wilson and Neville 2009). When stakeholders define or rename the object of evaluation, highlighting their strengths in addressing it, this reduces the negative labelling and opens the way for improvement.

Fuzzy cognitive mapping (FCM) allows groups or individuals to frame the evaluation problem in terms of their own understanding ot it (Andersson and Silver 2019). In a fuzzy cognitive mapping session, participants build a flexible causal model of how they see the problem by providing concepts and lexicon that are familiar to them. They describe the factors that influence the problem and discuss the directions and strengths of each relationship that impacts the problem. This mapping is called “fuzzy” because it assesses the relative influence of each relationship in the map: participants are asked to rate this influence on a scale of 1 (weak) to 5 (strong). This approach enriches the evaluation but also changes its appropriation.

FCM supports cultural safety in several ways. Requiring no literacy or language skills, it promotes equity and inclusion (Andersson and Silver 2019). It reduces the stigma or shame that individuals or groups who belong to previously excluded groups may feel. In FCM sessions, groups are organized by gender, age and type of stakeholder (e.g. patients, providers, programme managers) to ensure that all voices are represented. Facilitators meet with each group separately, reducing power imbalances between groups. Ideally, facilitators should be of the same gender and age and share the language and socio-cultural realities of the participants to reduce hierarchy with them. Facilitators are trained to reduce bias.

The results of the fuzzy cognitive map can inform the evaluation beyond problem formulation. It allows for stakeholder input into the questionnaire design, combining existing literature with stakeholders’ views and understandings of the mechanisms of change (Dion et al. 2019; Dion et al. 2022). Giles and colleagues used the tool to capture how a Mohawk community understands the factors influencing diabetes (Giles et al. 2007). Sarmiento and colleagues used the tool to explore how indigenous communities in Guerrero State perceive factors that influence maternal health to better design interventions (Ivan Sarmiento, Paredes-Solís, et al. 2020; Iván Sarmiento, Zuluaga, et al. 2020).


Institutional ethics boards and evaluation ethics committees almost invariably rely on Western epistemologies, expecting all aspects of the evaluation to be clarified before the evaluation begins (Cameron et al. 2010). Yet cultural safety involves participant input into evaluation protocols and tools, which is not usually possible before the evaluation begins. Culturally safe evaluations should also seek the approval of Aboriginal committees whenever possible and respect the ethical guidelines for Aboriginal research. In addition, they should apply the principles of ownership, control, access and possession regarding how “data and information from Indigenous peoples will be collected, protected, used and shared. These principles aim to enhance the information governance of Indigenous peoples” (Nations 1998).

Research design

Cultural safety can be applied to qualitative, quantitative and mixed methods evaluations. For example, Andersson and colleagues used a randomized controlled trial to evaluate local interventions to reduce domestic violence in partnership with 12 Aboriginal women’s shelters in Canada (Andersson et al. 2010). The shelter directors requested the randomized controlled trial. They felt the need to show the impact of their programme to apply for more funding (Andersson et al. 2010).

Development of data collection instruments and methods

Evaluation teams are often encouraged to use standardized questionnaires to benefit from their validity and reliability. Many conventional questionnaires focus on risk factors, and deficits rather than the strengths and resilience that characterize many indigenous and minority group worldviews. Cultural safety requires more flexibility, particularly in the variables included in the questionnaires. This means including factors that stakeholders perceive as important, even if they are not part of a standardized questionnaire. Very often, it is possible to define the themes of a questionnaire through a participatory process, for example, the fuzzy cognitive mapping mentioned above.

Data collection in culturally safe evaluations can be quantitative or qualitative. Each method can introduce substantial bias when controlled too closely by the evaluation team. Careful recruitment and training of a local data collection team who share a similar social and cultural context with the community may be a better strategy. Not only does this increase acceptability and response rates, but the evaluation promotes skill development in the community.

Data analysis and interpretation

Whether using qualitative or quantitative methods, the analysis and interpretation of data should reflect the worldviews of indigenous and minority groups. For example, in inductive content analysis, there are participatory categorisation and coding options (Liebenberg, Jamal, and Ikeda 2020). Even when the analysis is computerised, which makes wider participation difficult, frequent checking by community members and the separation of analysis (data analysis) from interpretation (what the results mean) help to support the voice of participants and increase the accuracy and relevance of the analysis.

Participants’ voices also play a potential role in formal statistical analysis. Andersson and colleagues use the weights generated by fuzzy cognitive mapping as Bayesian a priori, incorporating pre-existing beliefs and knowledge as a prior probability distribution. This allows the integration of the indigenous perspective into statistical analysis.

Communication and knowledge transfer activities

Dissemination and transfer of knowledge are essential phases in evaluation. At these stages, there is a high risk of exploiting participants and using the results to project a situation that the communities could not handle. Instead of separating knowledge transfer activities and treating them as the final product of the evaluation, a culturally safe evaluation integrates them into the evaluation process (Kothari, McCutcheon, and Graham 2017). This approach involves all stakeholders – programme staff, funders, participants and, where possible, policymakers – in the evaluation from the outset. All are thus involved in the design, implementation and interpretation. Andersson and colleagues have developed a protocol called SEPA (socialising evidence for participatory action) that allows these steps to be integrated into the research. The SEPA protocol involves stakeholders in the production of evidence; it also implies presenting the research data so that they can participate not only in its interpretation but also in developing solutions in dialogues. Solutions are thus contextualised, implemented and evaluated (Ledogar et al. 2017).

II. How is this approach useful for policy evaluation?

Cultural safety is the first question that should be asked when evaluating a given programme. In addition to answering ex-ante and ex-post evaluation questions, the cultural safety approach helps to amplify the voices of participants and beneficiaries. It should be a requirement for all evaluations of public services.  For example, in programmes aimed at addressing the needs of Aboriginal or minority people, involving them in the evaluation, hearing how they define the problem (what their needs are and how they perceive the relevance of the evaluation) and understanding how they perceive the cultural appropriateness of potential solutions can only strengthen the policy and make it more appropriate (Cram 2016; Cameron et al. 2010).

Cultural safety helps to avoid implementation barriers, increasing the programme’s effectiveness by including indigenous worldviews, needs and priorities (effectiveness evaluation). It can help to reduce inequalities (short-term impact). It can empower these communities and eventually lead to their self-determination (long-term impact).

III. An example of the use of this approach in reproductive health

A reproductive health project in a post-conflict region of northern Uganda used the cultural safety framework to improve reproductive health outcomes. The project received funding from a Canadian organisation (Belaid et al. 2020; Belaid et al. 2021).

The civil war (1986-2006) between the Lord’s Resistance Army (North) and the Museveni government (South) displaced more than 90% of the population in this region. This has increased long-standing tensions between northern and southern Uganda (Laruni 2015).

Northern communities are still recovering from the conflict. The region has low poverty, social opportunity and human development indicators (Esuruku 2019). The conflict has had a negative impact on health services, deteriorating maternal health (Chi et al. 2015b, 2015a). Women and girls in this region are less educated and poorer. They are much less likely to give birth in a health facility (Uganda Bureau of Statistics (UBOS) 2012).

Before launching the project, we invested time in building relationships with local stakeholders and developing networks to involve community members. Stakeholders included women and men of different ages, traditional midwives, service providers and government officials. We involved these groups in all activities of the programme design. Each group defined perinatal care outcomes according to their worldview. We used fuzzy cognitive mapping to collect and compare perspectives. Group discussions clarified the lexicon and cultural concepts associated with perinatal care. We used these concepts to design the questionnaire, as far as possible, also using standardised questions corresponding to the concepts identified by the stakeholders.

The groups met in a series of deliberative dialogues to discuss local evidence, generate lists of potential strategies for improving access to perinatal care and design a programme. We invited the groups to discuss who should deliver the programme, how and with what content. Participants identified several barriers to accessing perinatal care and proposed strategies to address the problems in a culturally safe way.

Cultural safety helped identify problems in perinatal service provision. Reflection on local evidence generated feasible community-led solutions. This, in turn, increased trust between community members and service providers.

IV. What are the criteria for judging the quality of the mobilisation of this approach?

Cultural safety can only be judged by programme beneficiaries (Wilson and Neville 2009; Cameron et al. 2010; CIET/PRAM 2022). However, evaluation teams can reflect on the following questions:

  • Do participants/beneficiaries report feeling culturally safe during the evaluation?
  • How will the intended beneficiaries be involved in each evaluation phase?
  • Do the terms of the evaluation lend themselves to partnership?
  • Does the evaluation build on the strengths of communities?
  • Does the evaluation increase ownership of the project or service and the evaluation products?
  • How are the methods adapted to the specific culture?
  • What is the anticipated impact of the evaluation on community self-determination?

V. What are the strengths and limitations of this approach compared to others?

A cultural safety framework has several advantages for evaluation teams and programme participants. It increases the local acceptability and relevance of the evaluation. It can guide the design of programmes and services, increasing their contextual appropriateness. In evaluating ongoing programmes, cultural safety is an interpretive lens for understanding how indigenous and minority communities experience these programmes and services. As an analytical lens, it can highlight how inequalities and social injustices are shaped, what changes are needed, and what barriers or facilitators to these changes are possible (Gerlach 2012).

The main challenge is to develop and agree on protocols for assessing the impact of cultural safety in the context of complex outcomes (Gerlach 2012; Tremblay et al. 2020). Cultural safety depends largely on each local context, as each cultural group is different and has its own way of seeing things and its own degree of adaptation to dominant representations (Cameron et al., 2010). However, as more and more evaluations apply a cultural safety framework, our experiences of best practices will accumulate and contribute to developing guidelines with a wide range of transferability.

Some bibliographical references to go further

Andersson, Neil. and Shea, Beverley. and Amaratunga, Carol. and McGuire, Patricia. and Sioui, Georges. 2010. “Rebuilding from Resilience: Research Framework for a Randomized Controlled Trial of Community-led Interventions to Prevent Domestic Violence in Aboriginal Communities.” Pimatisiwin, 8(2): 61-88.

Andersson, Neil. and Silver, Hilah. 2019. “Fuzzy cognitive mapping: An old tool with new uses in nursing research.” Journal of Advanced Nursing, 75, no.12: 3823-30.

Belaid, Loubna. and Atim, Pamela. and Atim, Eunice. and Ochola, Emmanuel. and Bayo, Pontius. and Oola, Janet. and Sarmiento, Ivan. and Zarowsky, Christina. and Andersson, Neil. 2020. “Marginalized women and services providers improve access to perinatal care in post-conflict Northern Uganda: socializing evidence for participatory action”. Family Medicine and Community Health 9:e000610.

Cameron, Mary. and Andersson, Neil. and McDowell, Ian. and Ledogar, Robert. 2010. “Culturally Safe Epidemiology: Oxymoron or Scientific Imperative.” Pimatisiwin, 8(2): 89-116.

Curtis, Elana. and Jones, Rhys. and Tipene-Leach, David. and Walker, Curtis. and Loring, Belinda. and Paine, Sarah-Jane. and Reid, Papaarangi. 2019. “Why cultural safety rather than cultural competency is required to achieve health equity: a literature review and recommended definition.” International Journal for Equity in Health, 18(1): 174. https://doi.org/10.1186/s12939-019-1082-3. https://doi.org/10.1186/s12939-019-1082-3.

Dion, Anna. and Carini-Gutierrez, Alessandro. and Jimenez, Vania. and Ben Ameur, Amal. and Robert, Emilie. and Joseph, Lawrence. and Andersson, Neil. 2022. “Weight of Evidence: Participatory Methods and Bayesian Updating to Contextualize Evidence Synthesis in Stakeholders’ Knowledge.” J Mix Methods Res, 16(3): 281-306. https://doi.org/10.1177/15586898211037412.

Gerlach, Alison J. 2012. “A Critical Reflection on the Concept of Cultural Safety.” Canadian Journal of Occupational Therapy, 79(3): 151-58. https://doi.org/10.2182/cjot.2012.79.3.4. https://journals.sagepub.com/doi/abs/10.2182/cjot.2012.79.3.4.

Giles, Brian. and Findlay, C. Scott. and Haas, George. and LaFrance, Brenda. and Laughing, Wesley. and Pembleton, Sakakohe. 2007. “Integrating conventional science and aboriginal perspectives on diabetes using fuzzy cognitive maps.” Social Science & Medicine, 64(3): 562-76. https://doi.org/https://doi.org/10.1016/j.socscimed.2006.09.007. http://www.sciencedirect.com/science/article/pii/S0277953606004758.

Gollan, Sharon. and Stacey, Kathleen. 2021. Australian Evaluation Society First Nations Cultural Safety Framework. Australian Evaluation Society (Melbourne, Australia).

Gomez Cardona, Liliana. and Brown, Kristyn. and McComber, Mary. and Outerbridge, Joy. and Parent-Racine, Echo. and Phillips, Allyson. and Boyer, Cyndy. and Martin, Codey. and  Splicer, Brooke. and Thompson, Darrell. and Yang, Michelle. and Velupillai, Gajanan. and Laliberté, Arlène. and Haswell, Melissa. and Linnaranta, Outi. 2021. “Depression or resilience? A participatory study to identify an appropriate assessment tool with Kanien’kéha (Mohawk) and Inuit in Quebec.” Social Psychiatry and Psychiatric Epidemiology, 56(10): 1891-902. https://doi.org/10.1007/s00127-021-02057-1. https://doi.org/10.1007/s00127-021-02057-1.

Ledogar, Robert. and Arosteguí, Jorge. and Hernández-Alvarez, Carlos. and Morales-Perez, Arcadio. and Nava-Aguilera, Elizabeth. and Legorreta-Soberanis, José. and Suazo-Laguna, Harold. and Belli, Alejandro. and Laucirica, Jorge. and Coloma, Josefina. and Harris, Eva. and Andersson, Neil. 2017. “Mobilising communities for Aedes aegypti control: the SEPA approach.” BMC Public Health, 17(1): 403. https://doi.org/10.1186/s12889-017-4298-4. https://doi.org/10.1186/s12889-017-4298-4.

Nations, Le Centre de Gouvernance de l’Information des Premières. 1998. “Les principes de PCAP® des Premières Nations.” Le Centre de Gouvernance de l’Information des Premières Nations.

Papps, Elaine. and Ramsden, Irihapeti. 1996. “Cultural Safety in Nursing: the New Zealand Experience.” International journal for quality in health care: journal of the International Society for Quality in Health Care / ISQua, 8: 491-7. https://doi.org/10.1093/intqhc/8.5.491.

Sarmiento, Ivan. and Paredes-Solís, Sergio. and Loutfi, David. and Dion, Anna. and Cockcroft, Anne. and Andersson, Neil. 2020. “Fuzzy cognitive mapping and soft models of indigenous knowledge on maternal health in Guerrero, Mexico.” BMC Medical Research Methodology, 20(1): 125. https://doi.org/10.1186/s12874-020-00998-w. https://doi.org/10.1186/s12874-020-00998-w.

Smith, Linda Tuhiwai. 2012. Decolonising methodologies. Research and Indigenous Peoples. New York, USA: Bloomsbury.

Smye, Vicky. and Browne, Annette. 2002. “‘Cultural safety’ and the analysis of health policy affecting aboriginal people.” Nurse Res, 9(3): 42-56. https://doi.org/10.7748/nr2002.

Williams, Robyn. 1999. “Cultural safety — what does it mean for our work practice?” Australian and New Zealand Journal of Public Health, 23(2): 213-14. https://doi.org/https://doi.org/10.1111/j.1467-842X.1999.tb01240.x. https://onlinelibrary.wiley.com/doi/abs/10.1111/j.1467-842X.1999.tb01240.x.

Wilson, Denise. and Neville, Stephen. 2009. “Culturally safe research with vulnerable populations.” Contemporary Nurse, 33(1): 69-79. https://doi.org/10.5172/conu.33.1.69. https://doi.org/10.5172/conu.33.1.69.


Icon for the Creative Commons Attribution-ShareAlike 4.0 International License

Policy Evaluation: Methods and Approaches Copyright © by Loubna Belaid and Neil Andersson is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License, except where otherwise noted.

Share This Book